Create Account

How to use R to parse FHM6 save files
#1

Want to dabble in statistical analysis?
Have you ever wanted to learn how to analyze the data in Franchise Hockey Manager 6? Look no further than this beginner tutorial based in R.


Code:
I don't know what type (if any) media this is, but I thought I would give this a chance.

Word count: 1967 without the code chunks

[Image: 8E70VfU.png]

[Image: canadice.gif]
Thanks to @sköldpaddor, @Ragnar, @Carpy48 and @High Stick King, for the signatures







[Image: jZtKPwK.png] | [Image: RyzkmSj.png] | [Image: HKi05IH.png]

[Image: EzY5jpl.png]
Reply
#2

This is entirely too smart for me

[Image: 66818_s.gif]
Reply
#3

Well done! I appreciate the work that goes into Rmd’s.

I might’ve missed something zooming through, but is your prediction based on the point estimate alone? Or are you doing something else?
Reply
#4
(This post was last modified: 05-05-2021, 01:46 PM by Canadice.)

05-05-2021, 01:40 PMGCool Wrote: Well done! I appreciate the work that goes into Rmd’s.

I might’ve missed something zooming through, but is your prediction based on the point estimate alone? Or are you doing something else?

In this tutorial the predictions are solely based on the point estimate. You can enter an interesting debate whether or not you are able to create confidence intervals for the mean number of wins, but I would argue that you cannot do that because the data is not independent. What I instead did was using the quantiles (25% and 75% percentiles) as some form of measurement of the uncertainty, and the more certain predictions are when the casino line falls under/over the upper/lower quantile.

[Image: 8E70VfU.png]

[Image: canadice.gif]
Thanks to @sköldpaddor, @Ragnar, @Carpy48 and @High Stick King, for the signatures







[Image: jZtKPwK.png] | [Image: RyzkmSj.png] | [Image: HKi05IH.png]

[Image: EzY5jpl.png]
Reply
#5

05-05-2021, 01:45 PMCanadice Wrote:
05-05-2021, 01:40 PMGCool Wrote: Well done! I appreciate the work that goes into Rmd’s.

I might’ve missed something zooming through, but is your prediction based on the point estimate alone? Or are you doing something else?

In this tutorial the predictions are solely based on the point estimate. You can enter an interesting debate whether or not you are able to create confidence intervals for the mean number of wins, but I would argue that you cannot do that because the data is not independent. What I instead did was using the quantiles (25% and 75% percentiles) as some form of measurement of the uncertainty, and the more certain predictions are when the casino line falls under/over the upper/lower quantile.

I support that. I guess simming in succession would not be independent, but re-loading the original file and running a single season again would be - or looping the 20-season simulation n=20, 30, etc. times. Either of those would be independent trials in my mind, but yeah. I like your method!
Reply
#6

05-05-2021, 01:49 PMGCool Wrote:
05-05-2021, 01:45 PMCanadice Wrote: In this tutorial the predictions are solely based on the point estimate. You can enter an interesting debate whether or not you are able to create confidence intervals for the mean number of wins, but I would argue that you cannot do that because the data is not independent. What I instead did was using the quantiles (25% and 75% percentiles) as some form of measurement of the uncertainty, and the more certain predictions are when the casino line falls under/over the upper/lower quantile.

I support that. I guess simming in succession would not be independent, but re-loading the original file and running a single season again would be - or looping the 20-season simulation n=20, 30, etc. times. Either of those would be independent trials in my mind, but yeah. I like your method!

That is actually what the tutorial assumes you do, simulate the entire season `nSims` number of times. I would still err on the side of caution, but then again the sims you do at the start of the season don't take into account player development or tactic changes during the season. So the conclusions made with or without a confidence interval might hold the same amount of truth in the end.

[Image: 8E70VfU.png]

[Image: canadice.gif]
Thanks to @sköldpaddor, @Ragnar, @Carpy48 and @High Stick King, for the signatures







[Image: jZtKPwK.png] | [Image: RyzkmSj.png] | [Image: HKi05IH.png]

[Image: EzY5jpl.png]
Reply
#7

05-05-2021, 01:56 PMCanadice Wrote:
05-05-2021, 01:49 PMGCool Wrote: I support that. I guess simming in succession would not be independent, but re-loading the original file and running a single season again would be - or looping the 20-season simulation n=20, 30, etc. times. Either of those would be independent trials in my mind, but yeah. I like your method!

That is actually what the tutorial assumes you do, simulate the entire season `nSims` number of times. I would still err on the side of caution, but then again the sims you do at the start of the season don't take into account player development or tactic changes during the season. So the conclusions made with or without a confidence interval might hold the same amount of truth in the end.

Precisely. It's a perfectly sound experiment with a little bit of questionable logic when making the inference. In my experience in school and work, those typically work out just fine. Haha
Reply
#8

The League needs more guides and walkthroughs for stuff like this. This is great work!

[Image: 59269_s.png]


S66 Damian Littleton


[Image: CsnVET2.png] || [Image: wu5MVvy.png]|| [Image: c8B2LE3.png]
Battleborn | Barracuda | Usa
Reply
#9

What is with all of these SHL people using R?

[Image: doubtfulalpha.gif]
[Image: 0XJkcN5.png]
[Image: sN8N4xa.png][Image: 639861613880541184.png] Cal Juice [Image: 639861613880541184.png][Image: RyzkmSj.png]
[Image: Eo2nBCt.png] Tomas Zadina
[Image: snacnei.png] Brady McIntyre
[Image: ice-level.svg]
Reply
#10
(This post was last modified: 05-05-2021, 05:28 PM by hotdog.)

05-05-2021, 04:18 PMcaltroit_red_flames Wrote: What is with all of these SHL people using R?

R is cool and good

[Image: ekovanotter.gif]
thanks @Carpy48 and @frithjofr and @rum_ham and @Julio Tokolosh and @Briedaqueduc for the sigs
Armada Inferno norway
Reply
#11

05-05-2021, 05:27 PMhotdog Wrote:
05-05-2021, 04:18 PMcaltroit_red_flames Wrote: What is with all of these SHL people using R?

R is cool and good

R is cool and good

[Image: 0XJkcN5.png]
Czechoslovakia PROFILE || UPDATE || RAGE. Rage 
[Image: luketd.gif]




Reply
#12

05-05-2021, 05:29 PMluketd Wrote:
05-05-2021, 05:27 PMhotdog Wrote: R is cool and good

R is cool and good
E is cool and good @geck

[Image: sIjpJeQ.png]





Reply
#13

When your real life industry awards you for your sales performance and then you log into your sim life and realize you aren’t too smart. LIFE.

[Image: Wally.png]






Reply




Users browsing this thread:
1 Guest(s)




Navigation

 

Extra Menu

 

About us

The Simulation Hockey League is a free online forums based sim league where you create your own fantasy hockey player. Join today and create your player, become a GM, get drafted, sign contracts, make trades and compete against hundreds of players from around the world.